Does the 38% rule work?

Partly. In our simulation, sitting 38% of the room’s length from the front wall gave more even bass than most positions, but 38% from the back wall (which the rule also allows) was one of the worst when the subwoofer was at the front. Treat it as a starting point, then listen or measure.

Do golden-ratio rooms sound better?

Not in our simulation. Rooms with golden-ratio proportions had bass no more even than random rooms of the same volume. Avoiding cubes and rooms whose dimensions are simple multiples of each other matters; the exact “ideal” ratio barely does.

Is it worth having two subwoofers?

For more than one seat, yes. Two subs on opposite walls reduced seat-to-seat bass variation by 32% compared with one sub in a corner, in line with Harman’s research.

Yes. The results are free to cite under CC BY 4.0 with a link to this page. The summary, every room’s results and the code are linked under “How we tested”.

Home Theater Rules of Thumb, Tested in 1,000 Simulated Rooms

Where to sit: the 38% rule and its rivals

For each of 1,000 random rooms, we placed a subwoofer at the front, then moved a seat along the room’s center line from 15% to 85% of the way back, and measured how even the bass was at each position from 25 to 120 Hz. Then we asked: where does each rule’s seat rank among those 71 positions?

Median percentile among 71 seat positions (shorter is better; dashed line = the middle, 50th percentile)

Subwoofer in a front cornerSubwoofer at the front-wall center

38% from the front 39th 24th

Middle of the room 55th 32nd

38% from the back 82nd 83rd

One-fifth from the back 42nd 20th

Against the back wall* 16th 3rd

Percentile of the bass-evenness score among positions from 15% to 85% of the room’s length; the dashed line is the 50th percentile. *Ears 0.35 m (about 14 in) from the back wall, as on a sofa pushed against it.

38% from the front wall ranked in the better half in both setups. 38% from the back wall (62% of the way back) did badly in both, even though the rule treats it as equivalent. The reason is the subwoofer: room modes are symmetrical, but with the sub at the front, the odd-numbered modes arrive at those two spots with opposite polarity, so they add up differently. The rule’s symmetry only holds if the source is too.

Against the back wall was the most even position of all, with no deep cancellations, but it got about 3.7 to 4.3 dB more bass than the average position along the room. That extra level is what people usually mean by boomy bass: heavy rather than missing notes. (Our score measures evenness and level, not how long the bass rings, so this is the louder half of “boomy”.) The best position varied a lot from room to room: with a corner sub, the most even seat was a median 27% of the way back, and in half the rooms between 23% and 78%.

Try it in your own roomThe subwoofer simulator uses the same model: move the seat and the sub and watch the bass change. Open →

Room proportions: golden ratios and “ideal” rooms

We rebuilt every room at the same volume with each famous set of proportions (height : width : length) and averaged the bass evenness over 16 seats in the middle of the room, with a corner subwoofer. Random rooms averaged 6.94 on this score, with the middle 80% between 6.57 and 7.33.

Bass evenness compared with a random room of the same volume (dB; left of center = better)

Within ±0.3 dB of random (shaded band)Worse* Outside the ITU-R BS.1116 proportion limits

Golden ratio 0.00, same as random

Louden −0.13

Bolt −0.14

Sepmeyer A * +0.05

Sepmeyer B −0.25

Sepmeyer C −0.07

Double square * +0.50

Cube * +1.90

The golden ratio, Louden and Bolt ratios pass; so did 56% of our random rooms. Bolt and Louden ratios as quoted in Meissner’s 2018 review; originals not inspected.

The “ideal” ratios were within a quarter of a decibel of a random room, smaller than the variation between random rooms and a tenth of the effect of where you sit. The golden ratio did no better than chance. What clearly does matter is avoiding cubes and double squares, where room modes stack on the same frequencies. Sepmeyer, whose ratios are among the most quoted, said himself that no clearly defined optimum emerged from his calculations.

One subwoofer or two

Six seats in two rows of three. Seat-to-seat variation: at each frequency, the standard deviation of the level across seats, averaged over 25–120 Hz. Evenness at each seat: the same score as above (lower is better).
Setup	Seat-to-seat variation	Evenness at each seat
One sub, front corner	4.51 dB	7.31
One sub, front-wall center	3.74 dB	7.92
Two subs, front and back wall centers	3.07 dB	7.42
Two subs, side wall centers	3.18 dB	9.63

Two subs on opposite walls made the bass far more consistent from seat to seat (32% less variation than a corner sub) while keeping each seat about as even as the best single position. That matches Harman’s 2006 research (Welti and Devantier), which found two well-placed subs nearly as effective as four. Side-wall pairs evened out the seats too, but made the response at each seat less even in our model.

The viewing and speaker rules: checked against the standards

Rule	What it works out to	Compared with
Sit 1.5 × the diagonal away	32° view	SMPTE reference distance: 31–33°
Sit 2 × the diagonal away	25° view	Below the 30° that most guidance treats as the low end
Samsung: diagonal ÷ 0.55	27° view	Relaxed TV watching
Samsung (buying guide): diagonal × 1.2	40° view	THX’s 40° cinema angle
Sony 4K minimum: 1.5 × picture height	61° view	The 61° maximum for 4K cited in CEDIA/CTA RP22; Sony calls it a minimum
The 83% rule (speaker spacing)	49° between speakers	Dolby: 44–60° for front left and right
Equilateral triangle	60° between speakers	ITU-R BS.1116 stereo reference; top of Dolby’s range
4/6/8 × image height (farthest viewer)	25° / 17° / 13°	Meeting-room legibility, not cinema immersion

The scorecard: every rule we checked

Half trueThe 38% ruleSit 38% of the room’s length from the front or back wall.Practitioner rule (Ethan Winer, RealTraps); origin not established

38% from the front ranked in the better half of seat positions (median 39th percentile, corner sub). 38% from the back was among the worst (82nd).
Mostly holdsRule of fifths (seat)Seat one-fifth of the room’s length from the back wall.Hi-fi rule of thumb (PS Audio); several variants

Better than average: 42nd percentile with a corner sub, 20th with a front-wall sub.
PartlyDon’t sit in the middleThe center of the room is a bass dead spot.Room-mode physics (first length mode)

With a corner sub, the middle was slightly worse than average (55th percentile) and 1 dB down on bass; with a front-wall sub, slightly better than average.
True, differentlyKeep the sofa off the back wallBass builds up at the back wall.CEDIA/CTA RP22; boundary physics

Against the back wall bass was the most even, but about 3.7–4.3 dB louder than the average position along the room: heavy, not hollow.
MythGolden-ratio roomA room proportioned 1 : 1.618 : 2.618 has the best bass.Audio folklore (golden cuboid)

Bass at the seats was no better than in a random room of the same volume (difference 0.00 dB).
Small effectSepmeyer, Louden & Bolt ratiosPublished “ideal” room ratios smooth the bass.Acoustics papers (1946–1971); Sepmeyer found no clear optimum

Slightly better than random rooms (up to 0.25 dB), far less than the effect of where you sit. Cubes (1.9 dB worse) and double squares (0.5 dB worse) really are bad.
HoldsTwo subs beat oneTwo subwoofers even out bass across seats.Welti & Devantier (Harman), 2006

Two subs on opposite walls cut seat-to-seat variation by 32% compared with one sub in a corner.

Show the 7 viewing-distance, speaker and decorating rules

Holds (low end)Sit 1.5× the diagonal awaySeating distance = 1.5 × screen size.HD-era retail rule (CNET quoted 1.5–2×)

1.5× gives a 32° view, right on SMPTE’s reference (31–33°). 2× gives 25°, small for 4K.
RelaxedSamsung: diagonal ÷ 0.55Viewing distance = diagonal ÷ 0.55.Samsung US support

Gives a 27° view: comfortable for TV, a little small for films.
HoldsThe 83% ruleSpeaker spacing = 0.83 × your distance to each speaker.Jim Smith, Get Better Sound

Works out to a 49° angle between the speakers, inside Dolby’s 44–60° for front left and right.
HoldsEquilateral triangleSpeakers and listener form an equilateral triangle.ITU-R BS.1116 stereo reference (60°)

A 60° pair: the reference for stereo listening rooms and the top of Dolby’s range for home theater fronts.
Wrong roomThe 4/6/8 ruleFarthest viewer at 4, 6 or 8 × the image height.Meeting-room display sizing (AVIXA-style legibility)

For reading content in meetings, not films: 4× height is a 25° view, 8× just 13°.
Not about TVsThe 3-5-7 ruleGroup things in threes, fives and sevens.Decorating folklore

A styling rule for arranging objects; nothing to do with screens or sound.
Styling onlyThe two-thirds sofa ruleSofa about two-thirds the width of the wall or rug.Interior-design proportion cue

A visual proportion, not a viewing or sound rule. Viewing distance and side-seat angles matter more.

How we tested

Rooms: 1,000 rectangular rooms with random dimensions: length 12–26 ft (front to back), width 10–20 ft (never wider than long), ceiling 7.5–10 ft. Seeded (seed 20261004), so the published code gives the same numbers every time. For the room-ratio test, each room was rebuilt at its own volume with each set of proportions, so ceiling height and seat positions scale with it.
Model: the standard modal model of a rectangular room with rigid walls (Kuttruff, Room Acoustics), summing every mode below 250 Hz with damping from a 0.5-second low-frequency decay time, plus the room’s pressure-chamber term. It’s the same model code as our subwoofer simulator, which sums modes to 320 Hz; the check below shows that makes no difference.
Positions: ears at 1.2 m (Dolby’s seated ear height); subwoofer drivers at 0.3 m, 0.3 m from the walls.
Score: the standard deviation of the level at the seat across 48 frequencies from 25 to 120 Hz. Lower means more even bass. It measures evenness, not loudness, which is why we report bass level separately for the back-wall seat.

Does it depend on our assumptions?

We reran the whole study changing one assumption at a time. The findings kept their direction and roughly their size in every run.

Percentiles: median rank among 71 seat positions with a corner subwoofer (lower is better). Room ratios: average unevenness compared with random rooms (higher is worse). Two subs: change in seat-to-seat variation against one corner sub. Back-wall bass: level against the average position along the room.
Run	38% from front	38% from back	Golden vs random	Cube vs random	Two subs	Back-wall bass
Published run (0.5 s decay, ears 1.2 m, sub 0.3 m up, modes to 250 Hz, 48 frequencies)	39th	82nd	0.00 dB	+1.9 dB	−32%	+3.7 dB
Livelier room: 0.7 s bass decay	39th	83rd	+0.15 dB	+1.9 dB	−35%	+4.1 dB
Deader room: 0.35 s bass decay	39th	79th	−0.11 dB	+1.7 dB	−28%	+3.1 dB
Lower ears: 1.0 m (deep sofa)	39th	83rd	+0.01 dB	+2.0 dB	−30%	+3.6 dB
Sub driver higher: 0.5 m	39th	82nd	−0.02 dB	+1.8 dB	−32%	+3.8 dB
Modes to 320 Hz (the simulator’s setting)	39th	82nd	−0.01 dB	+1.9 dB	−33%	+3.7 dB
Twice as many frequencies (96)	39th	82nd	0.00 dB	+1.9 dB	−32%	+3.7 dB

What the model leaves out. Real rooms have furniture, doors, openings, flexible walls and different materials, all of which shift room modes. The model is good at relative comparisons (this seat versus that one, one sub versus two) across many rooms; it doesn’t predict your room to the decibel. Rules about decoration and meeting-room displays weren’t simulated because there’s nothing acoustic to test.

Data and code: summary results (JSON, with the sensitivity runs), every room’s results (JSON, 0.9 MB), and the code that produced them: the study script and the room model (TypeScript). Data generated 2026-10-05. You’re welcome to cite or republish them under CC BY 4.0 with a link to this page.

We tested 14 home‑theater rules of thumb in 1,000 simulated rooms